Related
E+T Magazine

This article has been provided by Aly MacGregor, property executive at WSP

Have you heard the one about an AI that couldn’t create an image of a bike with square wheels, since all its training data had bikes with round wheels? You could take that as a humorous illustration of AI’s limitations – or a cautionary tale of the limitations of human engineers.

Most engineers I know take a degree of pleasure in demonstrating how clever they are, in this case outwitting AI. Yet, I question whether that’s the best demonstration of intelligence. After all, breaking a model isn’t necessarily that hard, if that’s your goal. We would be far smarter and better served to figure out new ways to use AI to address more meaningful questions.

Instead of figuring out how to break a model, engineers should be asking how to optimise design, improve efficiency, or solve complex challenges. For example, rather than ask about bikes with square wheels, we should be asking about practical challenges, such as creating...

Click here to read the full article

Parents

  • The problem I have with these AI systems is responsibility, the same as with self-driving cars. The article notes the problem that AI trusts the work of others but we do not know who the others are.

    “Since most current AI tools don’t get hung up on causality (the why) and can mine accumulated knowledge and previously developed solutions to present a range of “hows,” AI doesn’t have an ego that needs affirmation. The tool trusts the work of others (which we do need to check) and can provide impartial solutions.”

    If I have a design to calculate I can pick up my Engineers Handbook (which I can reference), select the formulas,  make the calculations showing my working and then sign it off. This is my responsibility.

    I can ask the question of some AI package and receive an answer. If I use this answer who carries the responsibility when an accident occurs? Is it me for being foolish enough to use an AI package or the AI package provider?

  • in reply to Roger B

    Hello Roger:

    One thing to note about your selection and use of formulas for design:- Most formulas are approximations and are only valid within predefined limits.

    The best engineering handbooks highlight these predefined limits, which the AI system does not usually identify.

    Peter Brooks

    Palm Bay FL.

Comment
  • in reply to Roger B

    Hello Roger:

    One thing to note about your selection and use of formulas for design:- Most formulas are approximations and are only valid within predefined limits.

    The best engineering handbooks highlight these predefined limits, which the AI system does not usually identify.

    Peter Brooks

    Palm Bay FL.

Children
No Data

Community Rules & Guidelines